Soil and Forest Structure Predicts Large-Scale Patterns of Occurrence and Local Abundance of a Widespread Amazonian Frog

Home , Allobates femoralis

Soil and forest structure predicts large-scale patterns of occurrence and local abundance of a widespread Amazonian frog

Anthony S. Ferreira1, Robert Jehle2,*, Adam J. Stow3,* and Albertina P. Lima4,* 1 Programa de Pós-Graduação em Ecologia, Instituto Nacional de Pesquisas da Amazônia, Manaus, Amazonas, Brazil 2 School of Environment and Life Sciences, University of Salford, Salford, UK 3 Department of Biological Sciences, Macquarie University, Sydney, NSW, Australia 4 Coordenação de Biodiversidade, Instituto Nacional de Pesquisas da Amazônia, Manaus, Amazonas, Brazil * These authors contributed equally to this work.

ABSTRACT The distribution of biodiversity within the Amazon basin is often structured by sharp environmental boundaries, such as large rivers. The Amazon region is also characterized by subtle environmental clines, but how they might affect the distributions and abundance of organisms has so far received less attention. Here, we test whether soil and forest characteristics are associated with the occurrence and relative abundance of the forest-ﬂoor dwelling Aromobatid frog, Allobates femoralis. We applied a structured sampling regime along an 880 km long transect through forest of different density. High detection probabilities were estimated for A. femoralis in each of the sampling modules. Using generalized linear mixed-effects models and simple linear regressions that take detectability into account, we show that A. femoralis is more abundant in open forests than in dense forests. The presence and relative abundance of A. femoralis is also positively Submitted 29 December 2017 associated with clay-rich soils, which are poorly drained and therefore likely Accepted 21 July 2018 support the standing water bodies required for reproduction. Taken together, we Published 9 August 2018 demonstrate that relatively easy-to-measure environmental features can explain the Corresponding author distribution and abundance of a widespread species at different spatial scales. Anthony S. Ferreira, [email protected] Such proxies are of clear value to ecologists and conservation managers working in large inaccessible areas such as the Amazon basin. Academic editor Leonardo Montagnani Additional Information and Subjects Biogeography, Conservation Biology, Ecology, Environmental Impacts, Population Declarations can be found on Biology page 15 Keywords Amazonia, Ecology, Allobates femoralis, Environmental heterogeneity, Ecological DOI 10.7717/peerj.5424 limiting factors, Tropical forest Copyright 2018 Ferreira et al. INTRODUCTION Distributed under The distribution of species is often fragmented, with favorable habitats being separated Creative Commons CC-BY 4.0 from each other by unsuitable habitats (Krebs, 1972; Hanski, 1999). Nevertheless, determining how habitat heterogeneity inﬂuences the distribution and abundance of

How to cite this article Ferreira et al. (2018), Soil and forest structure predicts large-scale patterns of occurrence and local abundance of a widespread Amazonian frog. PeerJ 6:e5424; DOI 10.7717/peerj.5424 species at various geographic scales remains one of the major challenges in ecology and conservation biology (Leibold et al., 2004; Fraterrigo, Wagner & Warren, 2004; Tews et al., 2004; McGarigal et al., 2016). Species interact with each other at fine scales, and habitat heterogeneity within a landscape moderates the broad-scale consistency of such interactions, producing variation in environmental effects at intermediate scales (Lawton, 1999). A persistent problem is explaining how abiotic and biotic factors affect the distributions of species across spatial scales which are hierarchical in nature (Fraterrigo, Wagner & Warren, 2004). Recent work has disentangled historical and environmental determinants for the spatial turnover of species assemblage compositions (Ricklefs & Schluter, 1993; Holyoak, Leibold & Holt, 2005; Bitar et al., 2017). However, for single species, the main focus has been on predicting entire ranges by extrapolating from local data on habitat requirements (e.g., using species distribution models, see Elith & Leathwick, 2009; Zurell et al., 2016), for which an understanding of the potential hierarchy of controls on species’ niches is critically important (Araujo & Luoto, 2007; Fraterrigo, Wagner & Warren, 2004). Studies that use empirical data to quantify habitat associations for single species across biogeographic scales, however, are surprisingly rare. In the Amazon basin, biogeographic and large-scale ecological studies are particularly challenging, due to difficulties of access and a general lack of baseline knowledge (Tuomisto et al., 2003; Laurance et al., 2004; Betts, Malhi & Roberts, 2008; Gardner et al., 2008; De Fraga et al., 2014). To explain patterns of endemism, marked ecological barriers such as large rivers have been found to predict the distribution of many taxa (Cracraft, 1985; Aleixo, 2006; Araripe et al., 2008; Ribas et al., 2012; Dias-Terceiro et al., 2015; Nazareno, Dick & Lohmann, 2017; Oliveira, Vasconcelos & Santos, 2017). While sharp environmental boundaries clearly influence the abundance and occurrence of organisms, the effect of more gradual clines of biotic and abiotic features are less appreciated (Quesada et al., 2012; Cintra et al., 2013; Emilio et al., 2013; Schietti et al., 2016). Under such a scenario, the abundance of organisms and patterns of local adaptation are shaped by continuous environmental changes across the landscape (Endler, 1977; Leite & Rogers, 2013; Dias-Terceiro et al., 2015; Bitar et al., 2017). Anurans are useful models to evaluate biogeographic and ecological determinants of species assemblages in tropical ecosystems owing to their high diversity, low vagility and specific environmental requirements (Zimmerman & Bierregaard, 1986; Ernst & Rödel, 2008; Menin et al., 2007; Keller et al., 2009). Furthermore, based on their life histories, groups of species can be assigned to specific guilds (e.g., lotic and lentic aquatic breeders, forest-floor dwellers and canopy species; for case studies from Amazonia see Zimmerman & Simberloff, 1996; Menin et al., 2007; Rojas-Ahumada, Landeiro & Menin, 2012; Landeiro, Waldez & Menin, 2014; Dias-Terceiro et al., 2015; Bitar et al., 2017). Related species may share behavioral, physiological, and morphological traits because of common ancestry, rather than as a result of being exposed to similar selection pressures and convergent evolution (Huey, 1987; Losos, 1990; Cadle & Greene, 1993; Zimmerman & Simberloff, 1996). Given that some habitat requirements are therefore likely to be shared by all individual species of a guild, it is remarkable that relatively little is known about habitat associations of particular species across significant parts of their range (but see e.g., Jorge et al., 2016).

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 2/26 Many anurans associate with standing water, so their populations are often patchily distributed across the landscape (Smith & Green, 2005). Therefore, a population size estimate is often unattainable as a metric for monitoring population status across large scales, due to its variability and work involved in collecting the data at each patch (Smith et al., 2014). Documenting patch (site) occupancy is a more practical option because it can be measured using presence/absence surveys, utilizing each site as a sampling unit (MacKenzie et al., 2002). Moreover, the identification of accessible and stable environmental features that reflect specific habitat requirements would circumvent the problem of directly measuring standing water bodies, which are often ephemeral, and therefore difficult to record. The present study identifies and characterizes important environmental parameters linked to the distribution and relative abundance of a widespread Amazonian forest-floor anuran, the Aromobatid frog Allobates femoralis, using a structured sampling regime spanning an 880 km environmental gradient across an interfluvial landscape. We focus on structural forest features and soil characteristics as surrogates for the species’ microhabitat requirements, and show that these features are able to predict both its regional occurrence as well as its large-scale relative abundance. Our findings suggest that the distribution and relative abundance of A. femoralis is shaped by gradual ecological clines. MATERIAL AND METHODS Study species The brilliant-thighed poison frog A. femoralis (Boulenger, 1883; Anura: Aromobatidae Grant et al., 2017; ♂ snout-vent length = 28–33 mm; ♀ snout-vent length = 33–35 mm) is widely distributed in non-flooded primary forests of the Amazon Basin and Guiana Shield in Brazil, Bolivia, Peru, Ecuador, Colombia, Guyana, Suriname, French Guiana and Venezuela (Lescure & Marty, 2000; Lima et al., 2006; Amézquita et al., 2009; Barrio-Amorós & Santos, 2010), although phylogeographic and taxonomic studies suggest that it comprises a suite of cryptic species (Grant et al., 2006, 2017; Fouquet et al., 2007; Santos et al., 2009; Simões, Lima & Farias, 2010). A. femoralis is active in leaf litter or on fallen tree trunks on the forest floor, with males exhibiting territorial behavior (Roithmair, 1992; Montanarin, Kaefer & Lima, 2011). Females lay eggs on dead leaves in male territories during the rainy season, and males use water bodies ranging from shed palm bracts, Brazil-nut capsules, puddles on fallen tree trunks, peccary wallows and temporary puddles on the forest floor to deposit tadpoles after hatching (Roithmair, 1994; Gascon, 1995; Ringler, Ursprung & Hödl, 2009; Beck, Thebpanya & Filiaggi, 2010; Ringler, Hödl & Ringler, 2015; Pašukonis et al., 2016, 2017). The availability and location of sites for tadpole deposition influences year-to-year displacement of individuals that survive more than one breeding season (Ringler, Ursprung & Hödl, 2009). The ephemeral occurrence of suitable bodies of water also sometimes forces male A. femoralis to deposit tadpoles more than 180 m away from their territories (Ringler et al., 2013), to which they reliably return (Pašukonis et al., 2013, 2014). Over the last two decades, A. femoralis has been used as a model species to address questions on diversification (e.g., Lougheed et al., 1999; Simões et al., 2008; Amézquita et al., 2009), sexual selection and

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 3/26 Figure 1 Purus-Madeira interfluve showing modules locations with the two main types of phytophysiognomy and the range of means of the relative abundance of A. femoralis. Purus-Madeira interfluve showing module locations (M1–M14), and the two main types of forest phytophysiognomy: lowland dense rainforest to the northeast (dark green) and lowland open rainforest to the southwest (light green). Red circles indicate the presence of A. femoralis with sizes representing relative abundances; blue circles indicate absence of A. femoralis. The scale of the symbols represents the range of means of the relative abundance of A. femoralis in the modules when present. The modules are approximately 50–60 km apart, with the exception of the distance between M11 and M12 which are separated by open transition forest and savanna. Map created in QGis 2.14 Essen (http://qgis.osgeo.org). Full-size  DOI: 10.7717/peerj.5424/fig-1

parental care (Ringler et al., 2015, 2016, 2017a; Ursprung et al., 2011; Pašukonis et al., 2016, 2017), movement ecology and spatial cognition (Pašukonis et al., 2016; Beck et al., 2017), and communication (Hödl, Amézquita & Narins, 2004; Amézquita, Castellanos & Hödl, 2005; Amézquita et al., 2006; Narins et al., 2005; Betancourth-Cundar et al., 2016; Ringler et al., 2017b).

Study area The Purus-Madeira interﬂuve (PMI) is located on the eastern boundary of the Inambari area of endemism in central-southern Amazonia, delimited by the Amazon, Purus and Madeira rivers and covering about 15.4 million hectares (Fearnside et al., 2009; Fig. 1).

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 4/26 It is of sedimentary origin in its northeastern parts (Late Pleistocene-Early Holocene see Sombroek, 2000; Rossetti, Toledo & Góes, 2005), where the water table is closer to the surface and large areas are waterlogged by temporary small streams during the rainy season (Fan & Miguez-Macho, 2010; Schietti et al., 2016). Soils are mainly plinthosols characterized by poor drainage; the predominant texture is silt and ﬁne sand in the northeast (Cintra et al., 2013; Martins et al., 2014) and podzolic soils with a predominant texture of clay and sand in the southwest (IBGE, 1997). The vegetation is classiﬁed as humid tropical lowland rainforest, composed of lowland dense rainforest in the northeast, with an about 40 m high canopy and frequent occurrence of palms in the understory, and lowland open rainforest with an about 40 m high canopy in the southeast (IBGE, 1997; Sousa, 2007). Considerable areas of savanna and transition between lowland open forest-savanna are present in the extreme southwest (Fig. 1). In the northeast of the PMI, forests are characterized by a lower basal area, lower canopy heights and lower mean wood density (lowland dense forests) compared to the central and southeast sites (lowland open forests), associated with an increase in rainfall seasonality and a change in soil structure (Sombroek, 2000; Cintra et al., 2013; Schietti et al., 2016). Rainfall is seasonal and heaviest between November and May. The mean annual precipitation ranges from 2,100 mm in the southwest to 2,800 mm in the northeast (Cintra et al., 2013; Alvares et al., 2014). Elevation ranges from 27 to 80 m above sea level (Sombroek, 2000). Temporary ponds occur in lower areas during the rainy season and are formed by undulating terrain (Rossetti, Toledo & Góes, 2005; Ferrão et al., 2018). A more detailed description of the study area is presented by Cintra et al. (2013) and Schietti et al. (2016).

Sampling design The BR-319 Highway has not been maintained as a thoroughfare since 1998, but still allows access to a large section of the PMI. The implementation of standardized sampling sites along this highway through the RAPELD (Rapid Assessment for Long Duration Ecological Projects; Magnusson et al., 2013) system has generated a large amount of environmental and biotic information (Levis et al., 2012; Baccaro et al., 2013; Emilio et al., 2013; Cintra et al., 2013; Martins et al., 2014; Schietti et al., 2016; Ferrão et al., 2016; 2018; De Abreu, Schietti & Anciães, 2018). For the present study, we collected data in 152 plots at 14 RAPELD research sites (modules, M) which are spread along a 880-km-long transect (Fig. 1; Table S1, see Magnusson et al., 2013 for more details). Standard RAPELD modules consist of two straight parallel ﬁve km long trails starting at one km distance from each other, with ﬁve plots of 250 Â 60 m (∼1.5 ha) that follow altitudinal isoclines to minimize within-plot environmental heterogeneity installed on each trail, at distances of one km (Fig. S1). A total of 11 modules were installed along the BR-319 Highway (M01–M11), with plots established at least one km from the road to avoid secondary forests. Three other modules are near the left bank of the Madeira River (M12–M14), with seven 250 m plots resulting in a total of 14 plots per module and installed in the same way as those along the BR-319 Highway. Detailed descriptions of RAPELD sampling units throughout the Amazon basin are available at https://ppbio.inpa.gov.br.

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 5/26 Allobates femoralis sampling We used time- and space-limited visual sampling (adapted from Crump & Scott, 1994; 2 min searches every 10 m along the 250 m long plots) and auditory searches (using playback to stimulate male responses) to quantify the relative abundance of A. femoralis. Each sampling session lasted about one hour along the 250 m long plot central line, and was undertaken by two experienced observers. The presence or absence of A. femoralis was recorded in segments of 10 m. As only a single record per segment was made, the maximum number of records was 25 per plot. We were careful not to record the same individual more than once on the same segment or in neighboring segments. Sampling was carried out during the breeding season, which coincides with the regional rainy season (Kaefer et al., 2012; Ferrão et al., 2018), between December and February 2010–2015. Data were collected during the daily periods of peak vocalization for the species (7:00–10:00 a.m. and 14:00–18:00 p.m., Kaefer et al., 2012). To determine if A. femoralis was present, we recorded data on calling males following the audio strip transect method outlined by Zimmerman (1994). The A. femoralis advertisement call is one of the best studied anuran vocalizations (Narins, Hödl & Grabul, 2003; Hödl, Amézquita & Narins, 2004; Amézquita, Castellanos & Hödl, 2005; Amézquita et al., 2006, 2009; Göd, Franz & Hödl, 2007; Simões et al., 2008; Ringler et al., 2017b), and in the study area is a trill composed of four whistle-like notes with ascending frequency modulation. Calling males of A. femoralis could be heard from a distance of 30 m, and are easily recognized. Each of the 152 plots was sampled once per ﬁeld season, giving a total of four surveys per site in total. To avoid temporal bias, two sampling sessions followed the directions M1–M14 and M14–M1, respectively, with the remaining two sampling sessions not undertaken in consecutive order. The research project was approved by the appropriate governmental bodies: Ministério do Meio ambiente (MMA), the Instituto Chico Mendes de Conservação da Biodiversidade (ICMBio license 13777) and the Sistema de Autorização e Informação em Biodiversidade (SISBIO license 7836-1) for the sampling of A. femoralis. All sampling procedures were approved by the ethics animal welfare committees of the Instituto Nacional de Pesquisas da Amazônia (CEUA/INPA: 041/2015) in accordance with established scientiﬁc practice guidelines and current Brazilian legislation.

Environmental variables To represent environmental effects on the distribution and relative abundance of A. femoralis along the PMI, we used physical soil parameters (sand, clay and silt contents) and forest structure (basal area and number of trees, see Table S2). In all plots, soil samples were extracted with an auger every 50 m along the central 250 m long transect to a depth of 10 cm (a total of six samples per plot). Samples were kept in sealed plastic bags for 2–5 days, air dried at ambient temperature, and mixed to form one composite sample per plot (Cintra et al., 2013). Soil physical structure was analyzed following a standard protocol of total dispersion, using sodium pyrophosphate to obtain relative clay, sand and silt contents (Donagema et al., 2011). Percentage of sand was determined with a 0.053 mm mesh sieve (tensile bolting cloth 16), dividing the remaining fraction into silt and clay

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 6/26 (Donagema et al., 2011). The proportion of clay was determined by separating particles of 20 mm from other particles, and the proportion of silt was determined by the difference between clay and sand values; for a full description of the methodology see Quesada et al. (2010) and Donagema et al. (2011). Forest structure was represented by the basal area and the hierarchical sum of the number of trees and palms in three size classes: (1) in a band (left side of the center line) of 250 Â 1 m (0.025 ha), counting all stems with diameter at breast height (DBH) 1 cm; (2) in a band of 250 Â 20 m (0.5 ha), counting all stems with DBH 10 cm; and (3) in a band of 250 Â 40 m (one ha), counting all stems with DBH 30 cm (Magnusson et al., 2005). Tree diameters were measured with a diametric tape to mm precision (Schietti et al., 2016). Total plot basal area was calculated by the sum of the transverse areas of all trees as π(DBH)2/4. Data for modules M1–11 were obtained from Schietti et al. (2016), whereas data for modules M12–14 were previously unpublished. Forest structure has previously been identiﬁed as an important factor affecting the distribution and abundance of frogs in Amazonia (Menin et al., 2007; Menin, Waldez & Lima, 2011; Landeiro, Waldez & Menin, 2014; Ferrão et al., 2018).

Data analysis To estimate occupancy and detection probabilities of A. femoralis for each module, we used a multi-season occupancy model based on four seasons of sampling and presence-absence data without covariates in the program PRESENCE v.12.10 (MacKenzie et al., 2003). Models developed for estimating occupancy can account for imperfect detection by using data from repeat surveys to discriminate between a species being either truly absent, or present but undetected (MacKenzie et al., 2003; 2006; Tyre et al., 2003). The probability of occupancy is only estimated for the first season in a multi-season analysis, with occupancy parameters for the subsequent seasons being derived using a recursive equation (MacKenzie et al., 2003). Detection probabilities might vary slightly among areas within each module as a function of change in habitat features. However, because we were interested in estimating detectability at the scale of tens of kilometres for each module, we used a model without covariates. We excluded from the model the three modules where A. femoralis has never been found (M3–M5). Taking detection probabilities per module into account, we used Generalized Linear Mixed-Effects Models (GLMMs) to investigate the variation in relative abundance along gradients with the forest-structure components (basal area and number of trees) and soil texture (sand, clay and silt) as fixed effects. Modules were included in the model as a random effect to account for the nested design (plots within modules, Zuur et al., 2009). We also used the detectability-corrected data per module to ran simple linear regressions to investigate relationships between the relative abundance of A. femoralis with each predictor environmental variable. We used Shapiro–Wilk analyses to test for significant deviation from normality, and Spearman’s coefficient to verify correlations between environmental variables. As M11 is characterized by high variation in silt across plots and a high abundance of A. femoralis in three plots appearing as an outlier, separate analyses with and without M11 were undertaken. As the number of records was

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 7/26 low in most plots, we used the sum of recorded individuals (instead of the mean) for the four samplings to represent the relative abundance of A. femoralis in each plot (following Bueno et al., 2012; Ferrão et al., 2018). To test for spatial autocorrelation among the modules, we used a Moran’s correlogram of geographical distance between pairs of modules and on the residuals of linear regression analysis of dependent variable between pairs of modules. Statistical analyses were carried out in the statistical platform R 3.2.3 (R Core Team, 2018). GLMM analyses were conducted with the packages lme4 (Bates, Maechler & Bolker, 2015), and the DHARMa package was used for the creation and simulation of scaled (quantile) residuals (Hartig, 2017). The marginal and conditional GLMMs r2 were calculated using the package MuMIn (Barton, 2018) and figures were compiled using the package visreg (Breheny & Burchett, 2017). We used the APE package to test for spatial auto-correlation (Paradis, Claude & Strimmer, 2004). We only show data in figures when the simple linear regressions model was significant at the 0.05 level. Maps were prepared with QGis 2.14 Essen (QGIS Development Team, 2016).

RESULTS We found A. femoralis in 11 of 14 modules. The average number of segments per module in which we detected A. femoralis varied from 1.5 to 10 in lowland dense rainforest modules when present, while in lowland open rainforest the mean the respective number ranged from 12 to 25 per module (Fig. 1). At the level of plots, the encountered number of A. femoralis was on average 80% higher in lowland open rainforest (mean = 15.67; SD = 17.21, Fig. 2) than in lowland dense rainforest (mean = 3; SD = 5.01, Fig. 2). Estimates from the multi-season model showed that the median proportion of plots predicted to be occupied per module was 0.44 (0.14–0.78), and that median A. femoralis detection probabilities were 0.78 (0.48–0.96) across the four surveys (Table 1). There was no correlation between the geographical distance among modules and their dissimilarity in soil structure (sand, clay and silt contents; Moran tests, p = 0.12, 0.39, and 0.80, respectively) or forest structure (basal area and number of trees; Moran tests, p = 0.70 and 0.40, respectively). The 880 km transect across the PMI was characterized by marked environmental - - gradients. Tree basal area ranged from 9.73 to 38.90 m2 ha 1. The number of trees per ha 1 varied from 1,335 to 11,475 considering all individuals with dbh 1cm(Figs. 3A and 3B), with more trees in the central area of the PMI and fewer trees towards marginal areas in the southwest (Fig. 3B). To the northeast of the PMI, the soil had high levels of silt (average 30–77%), while the soil to the southwest had high clay content (average 50–77%, Figs. 3C and 3D). Module 11 had the highest variation in silt between plots (22–70%, Fig. 3C), and a high relative abundance of A. femoralis. Exclusion of this module from the analyses greatly changed the slope, but not the direction, of the curve and masked the relationships with environmental variables. Therefore, we excluded this module from the simple linear regressions (module level), but not from the GLMMs (plot level). Four environmental variables were highly correlated (r > 0.6) with another variable: clay content with silt content, silt content and clay content with basal area, and basal

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 8/26 Figure 2 Median of A. femoralis relative abundance in the research modules along the Purus- Madeira interfluve. Median, quartiles and maximum and minimum values of A. femoralis relative abundance (sum of individuals per plot across all samplings taking detectability into account) in the research modules along the Purus-Madeira interfluve. Green bars (M1–M10) represent modules in the lowland dense rainforest, and orange bars (M11–M14) represent modules in the lowland open rainforest. Full-size  DOI: 10.7717/peerj.5424/fig-2

Table 1 Occupancy probabilities and detection of A. femoralis.

Modules O.M. SEOM p SEp M1 0.53 (0.00–1.00) 0.42 0.61 (0.00–1.00) 0.32 M2 0.28 (0.00–0.85) 0.29 0.48 (0.06–0.89) 0.21 M6 0.14 (0.00–0.41) 0.14 0.83 (0.52–1.00) 0.15 M7 0.21 (0.00–0.64) 0.20 0.56 (0.24–0.89) 0.17 M8 0.34 (0.01–0.68) 0.17 0.75 (0.51–1.00) 0.12 M9 0.44 (0.06–0.82) 0.19 0.81 (0.31–1.00) 0.25 M1O 0.57 (0.14–1.00) 0.22 0.90 (0.61–1.00) 0.15 M11 0.70 (0.42–0.99) 0.15 0.96 (0.89–1.00) 0.04 M12 0.78 (0.57–1.00) 0.11 0.91 (0.82–0.99) 0.04 M13 0.44 (0.10–0.78) 0.17 0.89 (0.74–1.00) 0.07 M14 0.42 (0.17–0.69) 0.13 0.87 (0.74–1.00) 0.07 Notes: Occupancy probabilities and detection of A. femoralis in 11 research sites which consist of two straight parallel 5 km long trails in central-southern Amazonia. O.M., proportion of plots predicted to be occupied in each module with confidence intervals (± 95%); SE, standard error; p, probability of detection for A. femoralis with confidence intervals (± 95%) for the four samplings in the Purus-Madeira interfluve.

area with number of trees (Table 2). Soils with more silt coincided with low relative abundance of A. femoralis (p = 0.006; Fig. 4A), and areas with lower tree basal area had a high relative abundance of A. femoralis (p = 0.05; Fig. 4B). No interaction was found between clay content and number of trees to determine the relative abundance of A. femoralis in the GLMMs (Figs. 4C and 4D). Detailed results of the GLMMs are presented in Table 3.

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 9/26 Figure 3 Median, quartiles and maximum and minimum values of environmental variables along the Purus-Madeira Interfluve. Median (black vertical line), quartiles and maximum and minimum values of (A) basal area; (B) number of trees; (C) silt content and (D) clay content across 14 research modules along the Purus-Madeira Interfluve. Full-size  DOI: 10.7717/peerj.5424/fig-3

Table 2 Pearson’s correlation coefficients between environmental variables along the Purus-Madeira interfluve. Variables Sand Silt Clay Basal area Number of trees Sand -0.30 -0.30 0.38 0.25 Silt -0.11 -0.92 0.77 0.54 Clay -0.39 -0.82 -0.86 -0.59 Basal area 0.30 0.64 -0.72 0.81 Number of trees 0.37 0.53 -0.54 0.74 Notes: Pearson’s correlation coefficients between environmental variables at the module level (upper right) and plot level (lower left) along the Purus-Madeira interfluve. Bold values correspond to correlated variables (r > 0.60, p < 0.05).

Simple linear regressions at the module level showed that the relative abundance of 2 A. femoralis was negatively related to silt content (F1.11 = 27.28, r = 0.69, p < 0.001; 2 Fig. 5A), basal area (F1.11 = 21.55, r = 0.63, p < 0.001; Fig. 5C) and number of trees 2 (F1.11 = 11.77, r = 0.47, p < 0.01; Fig. 5D), and positively related to clay content 2 (F1.11 = 24.78, r = 0.66, p < 0.001; Fig. 5B). Soil structure (silt and clay contents) explained up to 69% and 66%, and forest structure (basal area and number of trees) explained up to 63% and 47% of the variance in relative abundance of A. femoralis, respectively.

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 10/26 Figure 4 Partial regressions derived from generalized linear mixed-effects models. Partial regressions derived from generalized linear mixed-effects models (GLMMs) investigating the effects of (A) silt content, (B) basal area, (C) clay content and (D) number of trees per plot on A. femoralis relative abundance taking detectability into account in forests along the Purus-Madeira interfluve. Fitted lines indicate fixed-effect probabilities < 0.05 in the linear mixed-model analyses. Full-size  DOI: 10.7717/peerj.5424/fig-4

Table 3 Results of generalized linear mixed-effects models—GLMMs. 2 2 Dependent Fixed effects df AIC BIC logLik r marg r cond p variable Relative Silt + basal area 86 1,138.6 1,148.6 -565.3 0.06 0.47 0.006 0.05 abundance Clay + trees 86 1,147.9 1,157.9 -569.9 0.21 0.44 0.28 0.99 Sand + basal area 86 1,141.5 1,151.5 -566.7 0.02 0.49 0.09 0.05 Notes: Results of generalized linear mixed-effects models (GLMMs) for the relative abundance of A. femoralis taking detectability into account as a function of sand, clay and silt contents, basal area and number of trees (fixed effects). Modules was considered a random effect in all model. Marginal r2 values are for the models adjusted only considering fixed effects and the conditional r2 corresponds to the full model, including the random effect. The relative contribution of predictors is given by the standardized coefficients of the GLMMs. The probability for each predictor is shown in the sequence as they appear in the models. Standardized coefficients in bold have p < 0.05. The outlier module M11 is included in all models.

DISCUSSION We show that soil characteristics and forest structure can predict the distribution and relative abundance of the widespread forest-dwelling Amazonian frog A. femoralis at a range of spatial scales. The species is generally patchily distributed, and its occurrence and relative abundance is associated with gradual variation in environmental features. At a large geographic scale, we revealed that A. femoralis is more abundant in open forests and in areas with clay-rich soils. Due to the lack of RAPELD modules in the savanna-open

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 11/26 Figure 5 Relationship between mean A. femoralis relative abundance and silt and clay contents, basal area and number of trees. Relationship between mean A. femoralis relative abundance and (A) silt content, (B) clay content, (C) basal area and (D) number of trees per module, taking detectability into account and excluding the outlier M11 (open circles; see the main text for more details). The relationships were statistically signiﬁcant (percentage of silt: r2= 0.69, p< 0.001; percentage of clay: r2 = 0.66, p < 0.001; basal area: r2 = 0.63, p < 0.001 and number of trees: r2 = 0.47, p < 0.01). Full-size  DOI: 10.7717/peerj.5424/ﬁg-5

lowland forest transition area in the southwest of the PMI, this region remained unstudied. Dense forests are likely to inhibit dispersal, but the edges appear to provide suitable habitat for A. femoralis to reproduce, and these habitats may therefore act as corridors. Soil characteristics predict both forest structure (Cintra et al., 2013; Martins et al., 2014) and the presence of surface water (Woinarski, Fisher & Milne, 1999; Menin, Waldez & Lima, 2011) a requirement for reproduction by anurans with aquatic tadpoles such as A. femoralis. The physical composition of the soil is thus fundamentally interlinked with the occurrence of Amazonian frog species that depend on small stagnant water bodies on the forest ﬂoor. Our estimated detection probabilities demonstrate that multiple surveys of A. femoralis in the PMI are highly unlikely to result in false negatives (recorded absences when the species is actually present). We expected this result because A. femoralis is easily detectable

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 12/26 even when it is scarce, especially because males respond reliably to playback calls (Amézquita, Castellanos & Hödl, 2005; Amézquita et al., 2006; Betancourth-Cundar et al., 2016). Although the modules in the northeast (M1–M2 and M6–M9) show a more dispersed occupation of A. femoralis across plots, the average detectability did not markedly differ from modules of the southwest region (M10–M14) that have higher occurrences and relative abundances. In the Amazon basin, the minimum number of surveys necessary to determine the presence or absence of an amphibian species is not specified through general guidelines. As a consequence, for example environmental consultants regularly conduct only a single survey per locations, which can result in detection failure and therefore generate erroneous predictions of species’ responses to habitat factors (Gu & Swihart, 2003; Mazerolle et al., 2007). Our results suggest that the sampling regime applied for the present study (four surveys) is sufficient for the accurate detection of A. femoralis, although due to species-specific life histories this guide cannot be universally applied to other taxa (Smith et al., 2014). The use of environmental proxies for estimating the occupancy of particular species has received some criticism, especially regarding the lack of consideration of behavioral interactions and the relevance of spatial scale (Stephens et al., 2015). The data used to evaluate species distributions in predictive models are often spatially biased and rarely include abundance (Warren, 2012), and the effects of biotic interactions are expected to be averaged out at broader scales (Fraterrigo, Wagner & Warren, 2004). Our data are not subject to these problems because sampling was carried out in a spatially stratified and consistent manner using the RAPELD research modules. Furthermore, data were obtained at large enough spatial scales in a hierarchical framework to test whether soil and forest characteristics measured in situ act as proxies for explaining the distribution and relative abundance of A. femoralis across a substantial part of its distribution. Due to the unavailability of module-specific precipitation data we are unable to include information on rainfall in our models, although they would be useful to predict whether soil ditches and other structures suitable for pond formation are actually filled with water. Across the Amazon basin, differences in forest structure attributed to the physical characteristics of soil cause endogenous disturbances (Quesada et al., 2012; Cintra et al., 2013; Schietti et al., 2016), whereas edaphic or climatic factors can cause exogenous disturbances (Espírito-Santo et al., 2010, Cintra et al., 2013; Schietti et al., 2016). The two dominant forest phytophysiognomies in the PMI are spatially correlated with rainfall gradients (Sombroek, 2001), where forests in drier areas have lower stem densities and higher mass of individual trees compared to wetter forests, which have higher stem densities and lower individual tree mass (Cintra et al., 2013; Schietti et al., 2016). For A. femoralis, we show that these spatial differences in macro- and microhabitats shape its occurrence and relative abundance at both large and small scales, likely promoting its spatial differentiation. Parapatric segregation of populations associated with different forest formations has also recently been shown for an arboreal anuran occurring in the PMI (Osteocephalus taurinus, Ortiz, Lima & Werneck, 2018). The two main vegetation types in the PMI are related to geomorphology and the establishment of the current Amazonian drainage system (Rossetti, Toledo & Góes, 2005;

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 13/26 Hoorn et al., 2010; Latrubesse et al., 2010; Nogueira, Silveira & Guimarães, 2013). Tributaries in the dense northern forest drain into the Madeira River depression, while southern tributaries located within open forests drain into the Purus river depression (Fig. 1). According to Rossetti, Toledo & Góes (2005), seasonally flooded areas are linked to Holocene terrains, and the two major sedimentary units deposited 47,000 and 27,000 years ago are represented by lowland dense rainforest and lowland open rainforest. Drainage dynamics in each basin differ, and this coupled with different sedimentary loads and deposition age also influences vegetation establishment (Cohen et al., 2014). Old and well-drained soils are highly weathered (Emilio et al., 2013), whereas the soils in the northeast of the PMI are characterized by young and poorly drained sediments. Seasonally-flooded and poorly-structured soils provide a poorer substrate for root development and anchorage, resulting in frequent disturbances and high tree mortality rates, and consequently more dynamic, younger, and denser forests (Castilho et al., 2006; Feldpausch et al., 2011; Quesada et al., 2012; Cintra et al., 2013; Schietti et al., 2016). On the other hand, more structured soils with higher clay content are associated with older forests, where trees have higher individual mass and spacing between trees is larger (Castilho et al., 2006; Feldpausch et al., 2011; Emilio et al., 2013). Therefore, soil type can act as an environmental filter, selecting for different tree-growth strategies and partitioning the forest into patches of vegetation that are structurally distinct (Cintra et al., 2013; Emilio et al., 2013; Schietti et al., 2016). Soil type also reflects other attributes of the environment relevant to anurans. A study conducted in central Amazonia found higher production of litter in plane areas with clay-rich soils (Luizão et al., 2004). Other studies found higher species richness and abundance of anurans at sites with clay-rich soil, presumably linked to higher surface water availability (Hadden & Westbrooke, 1996; Woinarski, Fisher & Milne, 1999; Menin et al., 2007). Rain can create ponds isolated from streams on soil rich in clay (Menin, Waldez & Lima, 2011). Consequently, such soil is directly related to the availability of small standing water bodies (Menin, Waldez & Lima, 2011), a prerequisite for the reproduction of A. femoralis (Kaefer et al., 2012; Ringler, Hödl & Ringler, 2015). Gascon (1995) described the general advantages of natural pools for frogs, and Ringler, Hödl & Ringler (2015) demonstrated the impact of simulated peccary presence on A. femoralis population size, where installing artificial pools almost doubled the density of frogs within two years. Because assessing small water bodies that may serve as tadpole deposition sites is difficult over large areas, using proxies to predict the distribution and abundance of A. femoralis in environments such as the Amazon basin can save time and money. Clay-rich soils are also associated with a higher variety and density of terrestrial arthropods (Franklin, Magnusson & Luizão, 2005; Aguiar, Gualberto & Franklin, 2006), which generally provide food for leaf litter frogs. However, the relationships between invertebrates and soil texture have not been investigated in the PMI, and additional data are necessary to establish a link between prey density and A. femoralis occurrence and relative abundance. The influence of environmental heterogeneity on the distribution of frogs in the Amazonian lowlands generally depends on reproductive modes and breeding habitats

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 14/26 (Zimmerman & Bierregaard, 1986; Menin et al., 2007; Menin, Waldez & Lima, 2011; Landeiro, Waldez & Menin, 2014; Ferrão et al., 2018). A. femoralis is likely to have ecological requirements in common with other forest frogs with similar life histories. For example, the density of A. sumtuosus was positively related to the number of isolated pools on a local scale (Jorge et al., 2016), and the occurrence and relative abundance of tree frogs (Scinax), which use similar water bodies for reproduction, is also positively affected by soil silt content (Ferrão et al., 2018).

CONCLUSION Frogs face a range of threats, including habitat loss and fragmentation, disease, and introduced species (Kats & Ferrer, 2003; Stuart et al., 2004; Lips et al., 2006; Becker et al., 2007). Given the pace of development in the Amazon Basin and the limited resources available for conservation, quick and cost-effective methods for predicting anthropogenic impacts are required. Our assessment of A. femoralis revealed its patchy distribution, with higher occupancy probabilities in areas with high, open vegetation and clay-rich soil. Identifying relatively easy-to-measure environmental features that reﬂect the distribution and abundance of suits of organisms, including species-speciﬁc detection probabilities, are of clear value to ecologists and conservation managers.

ACKNOWLEDGEMENTS We thank Adrian A. Barnet for language editing of the ﬁrst manuscript version, William E. Magnusson and Fabrício Baccaro for help with statistical procedures, the Programa de Grande Escala da Biosfera-Atmosfera na Amazônia (LBA) for providing the ﬁeld assistant João Souza. We also like to thank Dr. Leonardo Montagnani and three anonymous referees for suggestions that have substantially improved the quality of the paper.

ADDITIONAL INFORMATION AND DECLARATIONS

Funding This work was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) who granted a PhD scholarship to Anthony S. Ferreira (number 161883/2014-1). Fieldwork and infrastructure were funded by the Instituto Nacional de Pesquisas da Amazônia (INPA), Programa de Suporte a Núcleos de Excelência (PRONEX, number 653/2009), CNPq—number 475559/2013-4, Programa de Pesquisa em Biodiversidade (PPBio/RAPELD), Centro Integrado de Estudos da Biodiversidade Amazônica (INCT—CENBAM) and Santo Antônio Energia S.A. This study was carried out in the modules along the Purus-Madeira interfluvial of Programa de Pesquisa em Longa Duração (PPBio/RAPELD) financed by FAPEAM and CNPq under grants conceded to Albertina P. Lima—number 653/2009. Collection of soil and forest structure data used in this study was funded by FAPESP/FAPEAM (465/2010) and additional funding was provided by PRONEX—FAPEAM (1600/2006) and PPBio Manaus (CNPq 558318/2009-6). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 15/26 Grant Disclosures The following grant information was disclosed by the authors: Conselho Nacional de Desenvolvimento Cientíﬁco e Tecnológico (CNPq). Instituto Nacional de Pesquisas da Amazônia (INPA). Programa de Suporte a Núcleos de Excelência: PRONEX, 653/2009. CNPq: 475559/2013-4. Programa de Pesquisa em Biodiversidade (PPBio/RAPELD). Centro Integrado de Estudos da Biodiversidade Amazônica (INCT—CENBAM). FAPEAM.

Competing Interests The authors declare that they have no competing interests.

Author Contributions Anthony S. Ferreira conceived and designed the experiments, performed the experiments, analyzed the data, contributed reagents/materials/analysis tools, prepared figures and/or tables, authored or reviewed drafts of the paper, approved the final draft. Robert Jehle authored or reviewed drafts of the paper, approved the final draft. Adam J. Stow authored or reviewed drafts of the paper, approved the final draft. Albertina P. Lima conceived and designed the experiments, performed the experiments, analyzed the data, contributed reagents/materials/analysis tools, authored or reviewed drafts of the paper, approved the final draft.

Animal Ethics The following information was supplied relating to ethical approvals (i.e., approving body and any reference numbers): All sampling procedures were discussed and approved by the ethics animal welfare committees of the Instituto Nacional de Pesquisas da Amazônia (CEUA/INPA: 041/2015) in accordance with good scientiﬁc practice guidelines and current Brazilian legislation.

Field Study Permissions The following information was supplied relating to ﬁeld study approvals (i.e., approving body and any reference numbers): The research project was approved by the appropriate governmental bodies: Ministério do Meio ambiente (MMA), through the Instituto Chico Mendes de Conservação da Biodiversidade (ICMBio license 13777) and Sistema de Autorização e Informação em Biodiversidade (SISBIO license 7836-1) for the sampling of A. femoralis. Data Availability The following information was supplied regarding data availability: The raw data are provided in the Supplemental File.

Supplemental Information Supplemental information for this article can be found online at http://dx.doi.org/10.7717/ peerj.5424#supplemental-information.

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 16/26 REFERENCES Aguiar NO, Gualberto TL, Franklin E. 2006. A medium-spatial scale distribution pattern of Pseudoscorpionida (Arachnida) in a gradient of topography (altitude and inclination), soil factors, and litter in a central Amazonia forest reserve, Brazil. Brazilian Journal Biology 66(3):791–802 DOI 10.1590/S1519-69842006000500004. Aleixo A. 2006. Historical diversification of floodplain forest specialist species in the Amazon: a case study with two species of the avian genus Xiphorhynchus (Aves: Dendrocolaptidae). Biological Journal of the Linnean Society 89(2):383–395 DOI 10.1111/j.1095-8312.2006.00703.x. Alvares CA, Stape JL, Sentelhas PC, Gonçalves JLM, Sparovek G. 2014. Köppen’s climate classification map for Brazil. Meteorologische Zeitschrift 22(6):711–728 DOI 10.1127/0941-2948/2013/0507. Amézquita A, Castellanos L, Hödl W. 2005. Auditory matching of male Epipedobates femoralis (Anura: Dendrobatidae) under field conditions. Animal Behaviour 70(6):1377–1386 DOI 10.1016/j.anbehav.2005.03.012. Amézquita A, Hödl W, Lima AP, Castellanos L, Erdtmann L, Araújo MC. 2006. Masking interference and the evolution of the acoustic communication system of the Amazonian poison frog Epipedobates femoralis. Evolution 60:1874–1887. Amézquita A, Lima AP, Jehle R, Castellanos L, Ramos O, Crawford AJ, Gasser H, Hödl W. 2009. Calls, colours, shapes, and genes: a multi-trait approach to the study of geographic variation in the Amazonian frog Allobates femoralis. Biological Journal of the Linnean Society 98(4):826–838 DOI 10.1111/j.1095-8312.2009.01324.x. Araripe J, Tagliaro CH, Rêgo PS, Sampaio I, Ferrari SF, Schneider H. 2008. Molecular phylogenetics of large-bodied tamarins, Saguinus spp. (Primates, Platyrrhini). Zoologica Scripta 37(5):461–467 DOI 10.1111/j.1463-6409.2008.00343.x. Araujo MB, Luoto M. 2007. The importance of biotic interactions for modelling species distributions under climate change. Global Ecology and Biogeography 16(6):743–753 DOI 10.1111/j.1466-8238.2007.00359.x. Baccaro FB, Rocha IF, Aguila BEG, Schietti J, Emilio T, Pinto JLPV, Lima AP, Magnusson WE. 2013. Changes in ground-dwelling ant functional diversity are correlated with water-table level in an Amazonian terra firme forest. Biotropica 45(6):755–763 DOI 10.1111/btp.12055. Barrio-Amorós CL, Santos JC. 2010. Amphibia, Anura, Dendrobatidae, Allobates femoralis (Boulenger, 1884): First confirmed country records, Venezuela. Check List 6(2):208–209 DOI 10.15560/6.2.208. Barton K. 2018. MuMIn: multi-model inference. R package version 1.40.4. Available at https:// CRAN.R-project.org/package=MuMIn. Bates D, Maechler M, Bolker B, Walker S. 2015. Fitting Linear Mixed-Effects Models using lme4. Journal of Statistical Software 67(1):1–48 DOI 10.18637/jss.v067.i01. Beck KB, Loretto M, Ringler M, Hödl W, Pašukonis A. 2017. Relying on known or exploring for new? Movement patterns and reproductive resource use in a tadpole-transporting frog. PeerJ 5:e3745 DOI 10.7717/peerj.3745. Beck H, Thebpanya P, Filiaggi M. 2010. Do Neotropical peccary species (Tayassuidae) function as ecosystem engineers for anurans? Journal of Tropical Ecology 26(4):407–414 DOI 10.1017/s0266467410000106. Becker CG, Fonseca CR, Haddad CFB, Batista RF, Prado PI. 2007. Habitat split and the global decline of amphibians. Science 318(5857):1775–1777 DOI 10.1126/science.1149374.

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 17/26 Betancourth-Cundar M, Lima AP, Hödl W, Amézquita A. 2016. Decoupled evolution between senders and receivers in the neotropical Allobates femoralis frog complex. PLOS ONE 11(6):e0155929 DOI 10.1371/journal.pone.0155929. Betts RA, Malhi Y, Roberts JT. 2008. The future of the Amazon: new perspectives from climate, ecosystem and social sciences. Philosophical Transactions of the Royal Society B: Biological Sciences 363(1498):1729–1735 DOI 10.1098/rstb.2008.0011. Bitar YOC, Tuomisto H, Pinheiro LC, Juen L, Santos-Costa MC. 2017. Species turnover in Amazonian frogs: low predict ability and large differences among forests. Biotropica 49(5):695–705 DOI 10.1111/btp.12428. Boulenger GA. 1883. On a collection of frogs from Yurimaguas, Huallaga River, Northern Peru. Proceedings of the Zoological Society of London 51(1):635–638 DOI 10.1111/j.1469-7998.1883.tb06669.x. Breheny P, Burchett W. 2017. visreg: visualization of regression models. R package Version 2.4-1. Available at https://CRAN.R-project.org/package=visreg. Bueno AS, Bruno RS, Pimentel TP, Sanaiotti TM, Magnusson WE. 2012. The width of riparian habitats for understory birds in an Amazonian forest. Ecological Applications 22(2):722–734 DOI 10.1890/11-0789.1. Cadle JE, Greene HW. 1993. Phylogenetic patterns, biogeography, and the ecological structure of Neotropical snake assemblages. In: Ricklefsand RE, Schluter D, eds. Species Diversity in Ecological Communities. Chicago: University of Chicago Press, 281–293. Castilho CV, Magnusson WE, Araújo NRO, Luizão RCC, Luizão FJ, Lima AP, Higuchi N. 2006. Variation in aboveground tree live biomass in a central Amazonian forest: effects of soil and topography. Forest Ecology and Management 234(1–3):85–96 DOI 10.1016/j.foreco.2006.06.024. Cintra BBL, Schietti J, Emillio T, Martins D, Moulatlet G, Souza P, Levis C, Quesada CA, Schöngart J. 2013. Soil physical restrictions and hydrology regulate stand age and wood biomass turnover rates of Purus–Madeira interfluvial wetlands in Amazonia. Biogeosciences 10(11):7759–7774 DOI 10.5194/bg-10-7759-2013. Cohen MCL, Rossetti DF, Pessenda LCR, Friaes YS, Oliveira PE. 2014. Late Pleistocene glacial forest of Humaitá—Western Amazonia. Palaeogeography, Palaeoclimatology, Palaeoecology 415:37–47 DOI 10.1016/j.palaeo.2013.12.025. Cracraft J. 1985. Historical biogeography and patterns of differentiation within the South American avifauna: areas of endemism. Ornithological Monographs 36:49–84 DOI 10.2307/40168278. Crump ML, Scott NJ Jr. 1994. Visual encounter surveys. In: Heyer WR, Donnelly MA, McDiarmid RW, Hayek LC, Foster MS, eds. Measuring and Monitoring Biological Diversity: Standard Methods for Amphibians. Washington: Smithsonian Institution Press, 84–92. De Abreu FHT, Schietti J, Anciães M. 2018. Spatial and environmental correlates of intraspecific morphological variation in three species of passerine birds from the Purus–Madeira interfluvium, Central Amazonia. Evolutionary Ecology 32(2–3):191–214 DOI 10.1007/s10682-018-9929-4. De Fraga R, Stow AJ, Magnusson W, Lima AP. 2014. The costs of evaluating species densities and composition of snakes to assess development impacts in Amazonia. PLOS ONE 9(8):e105453 DOI 10.1371/journal.pone.0105453. Dias-Terceiro RG, Kaefer IL, Fraga R, Araújo MC, Simões PI, Lima AP. 2015. A matter of scale: historical and environmental factors structure anuran assemblages from the Upper Madeira River, Amazonia. Biotropica 47(2):259–266 DOI 10.1111/btp.12197.

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 18/26 Donagema GK, Campos DVB, Calderano SB, Teixeira WG, Viana JHM. 2011. Manual de Métodos de Análise de Solo. Rio de Janeiro: Embrapa solos. Elith J, Leathwick JR. 2009. Species distribution models: ecological explanation and prediction across space and time. Annual Review of Ecology, Evolution, and Systematics 40(1):677–697 DOI 10.1146/annurev.ecolsys.110308.120159. Emilio T, Quesada CA, Costa FRC, Magnusson WE, Schietti J, Feldpausch TR, Brienen RJW, Baker TW, Chave J, Álvarez E, Araújo A, Bánki O, Castilho CV, Honorio EN, Killeen TJ, Malhi Y, Mendoza EMO, Monteagudo A, Neill D, Parada JA, Peña-Cruz A, Ramirez-Angulo H, Schwarz M, Silveira M, Steege H, Terborgh JW, Thomas R, Torres-Lezama A, Vilanova E, Phillips OL. 2013. Soil physical conditions limit palm and tree basal area in Amazonian forests. Plant Ecology & Diversity 7(1–2):215–229 DOI 10.1080/17550874.2013.772257. Endler JA. 1977. Geographic Variation, Speciation, and Clines. Princeton: Princeton University Press. Ernst R, Rödel MO. 2008. Patterns of community composition in two tropical tree frog assemblages: separating spatial structure and environmental effects in disturbed and undisturbed forests. Journal of Tropical Ecology 24(2):111–120 DOI 10.1017/S0266467407004737. Espírito-Santo FDB, Keller M, Braswell B, Nelson BW, Frolking S, Vicente G. 2010. Storm intensity and old-growth forest disturbances in the Amazon region. Geophysical Research Letters 37(11):1–6 DOI 10.1029/2010GL043146. Fan Y, Miguez-Macho G. 2010. Potential groundwater contribution to Amazon evapotranspiration. Hydrology and Earth System Sciences 14(10):2039–2056 DOI 10.5194/hess-14-2039-2010. Fearnside PM, Graça PMLA, Keizer EWH, Maldonado FD, Barbosa RI, Nogueira EM. 2009. Modelagem de desmatamento e emissões de gases de efeito estufa na região sob inﬂuência da rodovia Manaus–Porto Velho (Br-319). Revista Brasileira de Meteorologia 24(2):208–233 DOI 10.1590/s0102-77862009000200009. Feldpausch TR, Banin L, Phillips OL, Baker TR, Lewis SL, Quesada CA, Affum Baffoe K, Arets EJMM, Berry NJ, Bird M, Brondizio ES, Camargo P, Chave J, Djagbletey G, Domingues TF, Drescher M, Fearnside PM, França MB, Fyllas NM, Lopez-Gonzalez G, Hladik A, Higuchi N, Hunter MO, Iida Y, Salim KA, Kassim AR, Keller M, Kemp J, King DA, Lovett JC, Marimon BS, Marimon-Junior BH, Lenza E, Marshall AR, Metcalfe DJ, Mitchard ETA, Moran EF, Nelson BW, Nilus R, Nogueira EM, Palace M, Patiño S, Peh KSH, Raventos MT, Reitsma JM, Saiz G, Schrodt F, Sonké B, Taedoumg HE, Tan S, White L, Woll H, Lloyd J. 2011. Height-diameter allometry of tropical forest trees. Biogeosciences 8(5):1081–1106 DOI 10.5194/bg-8-1081-2011. Ferrão M, Colatreli O, de Fraga R, Kaefer IL, Moravec J, Lima AP. 2016. High species richness of Scinax treefrogs (Hylidae) in a threatened Amazonian landscape revealed by an integrative approach. PLOS ONE 11(11):e0165679 DOI 10.1371/journal.pone.0165679. Ferrão M, Fraga R, Moravec J, Kaefer IL, Lima AP. 2018. A new species of Amazonian snouted treefrog (Hylidae: Scinax) with description of a novel species-habitat association for an aquatic breeding frog. PeerJ 6:e4321 DOI 10.7717/peerj.4321. Fouquet A, Gilles A, Vences M, Marty C, Blanc M, Gemmell NJ. 2007. Underestimation of species richness in Neotropical frogs revealed by mtDNA analyses. PLOS ONE 2(10):e1109 DOI 10.1371/journal.pone.0001109.

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 19/26 Franklin E, Magnusson WE, Luizão FJ. 2005. Relative effects of biotic and abiotic factors on the composition of soil invertebrates communities in an Amazonian savannah. Applied Soil Ecology 29(3):259–273 DOI 10.1016/j.apsoil.2004.12.004. Fraterrigo JM, Wagner S, Warren RJ. 2004. Local-scale biotic interactions embedded in macroscale climate drivers suggest Eltonian noise hypothesis distribution patterns for an invasive grass. Ecology Letters 17(11):1447–1454 DOI 10.1111/ele.12352. Gardner TA, Barlow J, Araujo IS, Avila-Pires TC, Bonaldo AB, Costa JE, Esposito MC, Ferreira LV, Hawes J, Hernandez MI, Hoogmoed MS, Leite RN, Lo-Man-Hung NF, Malcolm JR, Martins MB, Mestre LA, Miranda-Santos R, Overal WL, Parry L, Peters SL, Ribeiro-Junior MA, Da Silva MN, Da Silva Motta C, Peres CA. 2008. The cost-effectiveness of biodiversity surveys in tropical forests. Ecology Letters 11(2):139–150 DOI 10.1111/j.1461-0248.2007.01133.x. Gascon C. 1995. Tropical larval anuran fitness in the absence of direct effects of predation and competition. Ecology 76(7):2222–2229 DOI 10.2307/1941695. Göd M, Franz A, Hödl W. 2007. The influence of internote-interval variation of the advertisement call on the phonotactic behaviour in male Allobates femoralis (Dendrobatidae). Amphibia- Reptilia 28(2):227–234 DOI 10.1163/156853807780202468. Gu W, Swihart RK. 2003. Absent or undetected? Effects of non-detection of species occurance on wildlife-habitats models. Biological Conservation 116(2):195–203 DOI 10.1016/s0006-3207(03)00190-3. Grant T, Frost DR, Caldwell JP, Gagliardo R, Haddad CFB, Kok PJR, Means DB, Noonan BP, Schargel WE, Wheeler W. 2006. Phylogenetic systematics of dart poison frogs and their relatives (Anura: Athesphatanura: Dendrobatidae). American Museum of Natural History Research Library 299:1–262 DOI 10.5531/sd.sp.14. Grant T, Rada M, Anganoy-Criollo M, Batista A, Dias PH, Jeckel AM, Machado DJ, Rueda-Almonacid JV. 2017. Phylogenetic systematics of dart-poison frogs and their relatives revisited (Anura: Dendrobatoidea). South American Journal of Herpetology 12(s1):S1–S90 DOI 10.2994/SAJH-D-17-00017.1. Hadden SA, Westbrooke ME. 1996. Habitat relationships of the herpetofauna of remnant Buloke Woodlands of Wimmera Plains, Victoria. Wildlife Research 23(3):363–372 DOI 10.1071/wr9960363. Hanski I. 1999. Metapopulation Ecology. Oxford: Oxford University Press. Hartig F. 2017. DHARMa: residual diagnostics for hierarchical (multi-level/mixed) regression models. R package Version 0.1.5. Available at http://florianhartig.github.io/DHARMa/. Hödl W, Amézquita A, Narins P. 2004. The rôle of call frequency and the auditory papillae in phonotactic behavior in male dart-poison frogs Epipedobates femoralis (Dendrobatidae). Journal of Comparative Physiology A 190(10):823–829 DOI 10.1007/s00359-004-0536-1. Holyoak M, Leibold MA, Holt RD. 2005. Metacommunities: Spatial Dynamics and Ecological Communities. Chicago: University of Chicago Press. Hoorn C, Wesselingh FP, Steege H, Bermudez MA, Mora A, Sevink J, Sanmartín I, Sanchez-Meseguer A, Anderson CL, Figueiredo JP, Jaramillo C, Riff D, Negri FR, Hooghiemstra H, Lundberg L, Stadler T, Särkinen T, Antonelli A. 2010. Amazonia through time: Andean uplift, climate change, landscape evolution, and biodiversity. Science 330(6006):927–931 DOI 10.1126/science.1194585. Huey RB. 1987. Phylogeny, history, and the comparative method. In: Feder ME, Bennett AF, Burggren WW, Huey RB, eds. New Directions in Ecological Physiology. Cambridge: Cambridge University Press, 76–101.

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 20/26 IBGE. 1997. Recursos naturais e meio ambiente: uma visão do Brasil. Second Edition. Rio de Janeiro: Instituto Brasileiro de Geografia e Estatística (IBGE), 208. Jorge RF, Simões PI, Magnusson WE, Lima AP. 2016. Fine-scale habitat heterogeneity explains the local distribution of two Amazonian frog species of concern for conservation. Biotropica 48(5):697–703 DOI 10.1111/btp.12333. Kaefer IL, Montanarin A, Da Costa, Rosangela S, Lima AP. 2012. Temporal patterns of reproductive activity and site attachment of the brilliant-thighed frog Allobates femoralis from central Amazonia. Journal of Herpetology 46(4):549–554 DOI 10.1670/10-224. Kats LB, Ferrer RP. 2003. Alien predators and amphibian declines: review of two decades of science and the transition to conservation. Diversity and Distributions 9(2):99–110 DOI 10.1046/j.1472-4642.2003.00013.x. Keller A, Rödel MO, Linsenmair KE, Grafe TU. 2009. The importance of environmental heterogeneity for species diversity and assemblage structure in Bornean stream frogs. Journal of Animal Ecology 78(2):305–314 DOI 10.1111/j.1365-2656.2008.01457.x. Krebs CJ. 1972. Ecology: The Experimental Analysis of Distribution and Abundance. New York: Harper & Row Press, 694. Landeiro VL, Waldez F, Menin M. 2014. Spatial and environmental patterns of Amazonian anurans: Differences between assemblages with aquatic and terrestrial reproduction, and implications for conservation management. Natureza & Conservação 12(1):42–46 DOI 10.4322/natcon.2014.008. Latrubesse EM, Cozzuol M, Silva-Caminha SAF, Rigsby CA, Absy ML, Jaramillo C. 2010. The late Miocene paleogeography of the Amazon basin and the evolution of the Amazon river system. Earth-Science Reviews 99(3–4):99–124 DOI 10.1016/j.earscirev.2010.02.005. Laurance WF, Albernaz AKM, Fearnside PM, Vasconcelos HL, Ferreira LV. 2004. Deforestation in Amazonia. Science 304(5674):1109b–1111b DOI 10.1126/science.304.5674.1109b. Lawton JH. 1999. Are there general laws in ecology? Oikos 84:177–192 DOI 10.2307/3546712. Leibold MA, Holyoak M, Mouquet N, Amarasekare P, Chase JM, Hoopes MF, Holt RD, Shurin JB, Law R, Tilman D, Loreau M, Gonzalez A. 2004. The metacommunity concept: a framework for multi-scale community ecology. Ecology Letters 7(7):601–613 DOI 10.1111/j.1461-0248.2004.00608.x. Leite RN, Rogers DS. 2013. Revisiting Amazonian phylogeography: insights into diversification hypotheses and novel perspectives. Organisms Diversity & Evolution 13(4):639–664 DOI 10.1007/s13127-013-0140-8. Lescure J, Marty C. 2000. Atlas des amphibiens de Guyane. Patrimoine Naturel 45:1–288. Levis C, Souza PF, Schietti J, Emilio T, Pinto JLPV, Clement CR, Costa FRC. 2012. Historical human footprint on modern tree species composition in the Purus-Madeira Interfluve, Central Amazonia. PLOS ONE 7(11):e48559 DOI 10.1371/journal.pone.0048559. Lima AP, Magnuson WE, Menin M, Erdtmann LK, Rodrigues DJ, Keller C, Hödl W. 2006. Guia de sapos da Reserva Adolpho Ducke, Amazônia Central. Manaus: Áttema Design Editorial, 168. Lips KR, Brem F, Brenes R, Reeve JD, Alford RA, Voyles J, Carey C, Livo L, Pessier AP, Collins JP. 2006. Emerging infectious disease and the loss of biodiversity in a Neotropical amphibian community. Proceedings of the National Academy of Sciences of the United States of America 103(9):3165–3170 DOI 10.1073/pnas.0506889103. Losos JB. 1990. Ecomorphology, performance capability, and scaling of West Indian Anolis lizards: an evolutionary analysis. Ecological Monographs 60(3):369–388 DOI 10.2307/1943062.

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 21/26 Lougheed SC, Gascon C, Jones DA, Bogart JP, Boag PT. 1999. Ridges and rivers: a test of competing hypothesis of Amazonian diversification using a dart-poison frog (Epipedobates femoralis). Proceedings of the Royal Society B: Biological Sciences 266(1431):1829–1835 DOI 10.1098/rspb.1999.0853. Luizão RCC, Luizão FJ, Paiva RQ, Monteiro TF, Sousa LS, Kruijt B. 2004. Variation of carbon and nitrogen cycling processes along a topographic gradient in a central Amazonian forest. Global Change Biology 10(5):592–600 DOI 10.1111/j.1529-8817.2003.00757.x. MacKenzie DI, Nichols JD, Hines JE, Knutson MG, Franklin AB. 2003. Estimating site occupancy, colonization and local extinction when a species is detected imperfectly. Ecology 84(8):2200–2207 DOI 10.1890/02-3090. MacKenzie DI, Nichols JD, Lachman GB, Droege S, Royle JA, Langtimm CA. 2002. Estimating site occupancy rates when detection probabilities are less than one. Ecology 83(8):2248–2255 DOI 10.2307/3072056. MacKenzie DI, Nichols JD, Royle AR, Pollock KH, Bailey LL, Hines JE. 2006. Occupancy Estimation and Modeling: Inferring Patterns and Dynamics of Species Occurrence. Burlington: Elsevier/Academic Press, 324. Magnusson WE, Braga-Neto R, Pezzini F, Baccaro F, Bergallo H, Penha J, Rodrigues D, Lima AP, Albernaz A, Hero JM, Lawson B, Castilho C, Drucker C, Franklin E, Mendonça F, Costa F, Galdino G, Castley G, Zuanon J, Vale J, Santos JLC, Luizão R, Cintra R, Barcosa RI, Lisboa A, Koblitz RV, Cunha CN, Pontes ARM. 2013. Biodiversidade e monitoramento ambiental integrado: o sistema RAPELD na Amazoônia. Santo Andren: Attema. Magnusson WE, Lima AP, Luizão R, Luizão F, Costa FRC, Castilho CV, Kinupp V. 2005. RAPELD: a modification of the Gentry method for biodiversity surveys in long-term ecological research sites. Biota Neotropica 5(2):19–24 DOI 10.1590/s1676-06032005000300002. Martins DL, Schietti J, Feldpausch TR, Luizão FJ, Phillips OL, Andrade A, Castilho CV, Laurance SG, Oliveira A, Amaral IL, Toledo JJ, Lugli LF, Pinto JLPV, Mendoza EMO, Quesada CA. 2014. Soil-induced impacts on forest structure drive coarse woody debris stocks across central Amazonia. Plant Ecology & Diversity 8(2):229–241 DOI 10.1080/17550874.2013.879942. Mazerolle MJ, Bailey LL, Kendall WL, Royle JA, Converse SJ, Nichols JD. 2007. Making great leaps forward: accounting for detectability in herpetological field studies. Journal of Herpetology 41(4):672–689 DOI 10.1670/07-061.1. McGarigal K, Wan HY, Zeller KA, Timm BC, Cushman SA. 2016. Multi-scale habitat selection modeling: a review and outlook. Landscape Ecology 31(6):1161–1175 DOI 10.1007/s10980-016-0374-x. Menin M, Lima AP, Magnusson WE, Waldez F. 2007. Topographic and edaphic effects on the distribution of terrestrially reproducing anurans in Central Amazonia: mesoscale spatial patterns. Journal of Tropical Ecology 23(5):539–547 DOI 10.1017/S0266467407004269. Menin M, Waldez F, Lima AP. 2011. Effects of environmental and spatial factors on the distribution of anuran species with aquatic reproduction in central Amazonia. Herpetological Journal 21:255–261. Montanarin A, Kaefer IL, Lima AP. 2011. Courtship and mating behaviour of the brilliant- thighed frog Allobates femoralis from Central Amazonia: implications for the study of a species complex. Ethology Ecology & Evolution 23(2):141–150 DOI 10.1080/03949370.2011.554884. Narins PM, Hödl W, Grabul DS. 2003. Bimodal signal requisite for agonistic behavior in a dart-poison frog, Epipedobates femoralis. Proceedings of the National Academy of Sciences of the United States of America 100(2):577–580 DOI 10.1073/pnas.0237165100.

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 22/26 Narins PM, Grabul DS, Soma KK, Gaucher P, Hödl W. 2005. Cross-modal integration in a dart-poison frog. Proceedings of the National Academy of Sciences of the United States of America 102:2425–2429 DOI 10.1073/pnas.0406407102. Nazareno AG, Dick CW, Lohmann LG. 2017. Wide but not impermeable: Testing the riverine barrier hypothesis for an Amazonian plant species. Molecular Ecology 26(14):3636–3648 DOI 10.1111/mec.14142. Nogueira ACR, Silveira R, Guimarães JTF. 2013. Neogene–Quaternary sedimentary and paleovegetation history of the eastern Solimões Basin, central Amazon region. Journal of South American Earth Sciences 46:89–99 DOI 10.1016/j.jsames.2013.05.004. Oliveira U, Vasconcelos MF, Santos AJ. 2017. Biogeography of Amazon birds: rivers limit species composition, but not areas of endemism. Scientific Reports 7(1):29–92 DOI 10.1038/s41598-017-03098-w. Ortiz DA, Lima AP, Werneck FP. 2018. Environmental transition zone and rivers shape intraspecific population structure and genetic diversity of an Amazonian rain forest tree frog. Evolutionary Ecology 32(4):359–378 DOI 10.1007/s10682-018-9939-2. Paradis E, Claude J, Strimmer K. 2004. APE: analyses of phylogenetics and evolution in R language. Bioinformatics 20(2):289–290 DOI 10.1093/bioinformatics/btg412. Pašukonis A, Beck KB, Fischer MT, Weinlein S, Stückler S, Ringler E. 2017. Induced parental care in a poison frog: a tadpole cross-fostering experiment. Journal of Experimental Biology 220(21):3949–3954 DOI 10.1242/jeb.165126. Pašukonis A, Loretto MC, Landler L, Ringler M, Hödl W. 2014. Homing trajectories and initial orientation in a Neotropical territorial frog, Allobates femoralis (Dendrobatidae). Frontiers in Zoology 11(1):29 DOI 10.1186/1742-9994-11-29. Pašukonis A, Ringler M, Brandl HB, Mangione R, Ringler E, Hödl W. 2013. The homing frog: high homing performance in a territorial dendrobatid frog Allobates femoralis (Dendrobatidae). Ethology 119(9):762–768 DOI 10.1111/eth.12116. Pašukonis A, Trenkwalder K, Ringler M, Ringler E, Mangione R, Steininger J, Warrington I, Hödl W. 2016. The signifiance of spatial memory for water fiding in a tadpole-transporting frog. Animal Behaviour 116:89–98 DOI 10.1016/j.anbehav.2016.02.023. QGIS Development Team. 2016. QGIS Geographic Information System. Open Source Geospatial Foundation Project. Available at http://qgis.osgeo.org. Quesada CA, Lloyd J, Schwarz M, Patiño S, Baker TR, Cz- imczik C, Fyllas NM, Martinelli L, Nardoto GB, Schmer-ler J, Santos AJB, Hodnett MG, Herrera R, Luizão FJ, Arneth A, Lloyd G, Dezzeo N, Hilke I, Kuhlmann I, Raessler M, Brand WA, Geilmann H, Moraes Filho JO, Carvalho FP, Araujo Filho RN, Chaves JE, Cruz Junior OF, Pimentel TP, Paiva R. 2010. Variations in chemical and physical properties of Amazon forest soils in relation to their genesis. Biogeosciences 7(5):1515–1541 DOI 10.5194/bg-7-1515-2010. Quesada CA, Phillips OL, Schwarz M, Czimczik CI, Baker TR, Patiño S, Fyllas NM, Hodnett MG, Herrera R, Almeida S, Alvarez Dávila E, Arneth A, Arroyo L, Chao KJ, Dezzeo N, Erwin T, di Fiore A, Higuchi N, Honorio Coronado E, Jimenez EM, Killeen T, Lezama AT, Lloyd G, López-González G, Luizão FJ, Malhi Y, Monteagudo A, Neill DA, Núñez Vargas P, Paiva R, Peacock J, Peñuela MC, Peña Cruz A, Pitman N, Priante Filho N, Prieto A, Ramírez H, Rudas A, Salomão R, Santos AJB, Schmerler J, Silva N, Silveira M, Vásquez R, Vieira I, Terborgh J, Lloyd J. 2012. Basin-wide variations in Amazon forest structure and function are mediated by both soils and climate. Biogeosciences 9(6):2203–2246 DOI 10.5194/bg-9-2203-2012.

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 23/26 R Core Team. 2018. R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing. Available at https://www.R-project.org/. Ribas CC, Aleixo A, Nogueira ACR, Miyaki CY, Cracraft JA. 2012. Palaeobiogeographic model for biotic diversification within Amazonia over the past three million years. Proceedings of the Royal Society B: Biological Sciences 279(1729):681–689 DOI 10.1098/rspb.2011.1120. Ricklefs RE, Schluter D. 1993. Species diversity in ecological communities: historical and geographic perspectives. Chicago: University of Chicago Press, 424. Ringler E, Beck KB, Weinlein S, Huber L, Ringler M. 2017a. Adopt, ignore, or kill? Male poison frogs adjust parental decisions according to their territorial status. Scientific Reports 7(1):43544 DOI 10.1038/srep43544. Ringler M, Hödl W, Ringler E. 2015. Populations, pools, and peccaries: simulating the impact of ecosystem engineers on rainforest frogs. Behavioral Ecology 26(2):340–349 DOI 10.1093/beheco/aru243. Ringler M, Szipl G, Hödl W, Khil L, Kofler B, Lonauer M, Provin C, Ringler E. 2017b. Acoustic ranging in poison frogs—it is not about signal amplitude alone. Behavioral Ecology and Sociobiology 71(8):114 DOI 10.1007/s00265-017-2340-2. Ringler E, Pašukonis A, Fitch WT, Huber L, Hödl W, Ringler M. 2015. Flexible compensation of lost uniparental care in a poison frog. Behavioral Ecology 26(4):1219–1225 DOI 10.1093/beheco/arv069. Ringler E, Pašukonis A, Ringler M, Huber L. 2016. Sex-specific offspring discrimination reflects respective risks and costs of misdirected care in a poison frog. Animal Behaviour 114:173–179 DOI 10.1016/j.anbehav.2016.02.008. Ringler E, Pašukonis A, Hödl W, Ringler M. 2013. Tadpole transport logistics in a Neotropical poison frog: indications for strategic planning and adaptive plasticity in anuran parental care. Frontiers in Zoology 10(1):67 DOI 10.1186/1742-9994-10-67. Ringler M, Ursprung E, Hödl W. 2009. Site fidelity and patterns of short- and long-term movement in the brilliant-thighed poison frog Allobates femoralis (Aromobatidae). Behavioral Ecology and Sociobiology 63(9):1281–1293 DOI 10.1007/s00265-009-0793-7. Roithmair ME. 1992. Territoriality and male mating success in the Dart-poison Frog, Epipedobates femoralis (Dendrobatidae, Anura). Ethology 92(4):331–343 DOI 10.1111/j.1439-0310.1992.tb00970.x. Roithmair ME. 1994. Field studies on reproductive behaviour in two dart-poison frog species (Epipedobates femoralis, Epipedobates trivittatus) in Amazonian Peru. Herpetological Journal 4:77–85. Rojas-Ahumada DP, Landeiro VL, Menin M. 2012. Role of environmental and spatial processes in structuring anuran communities across a tropical rain forest. Austral Ecology 37(8):865–873 DOI 10.1111/j.1442-9993.2011.02330.x. Rossetti DF, Toledo PM, Góes AM. 2005. New geological framework for Western Amazônia (Brazil) and implications for biogeography and evolution. Quarternary Research 63(1):78–89 DOI 10.1016/j.yqres.2004.10.001. Santos JC, Coloma LA, Summers K, Caldwell JP, Ree R, Cannatella DC. 2009. Amazonian amphibian diversity is primarily derived from Late Miocene Andean lineages. PLOS Biology 7(3):e1000056 DOI 10.1371/journal.pbio.1000056. Schietti J, Martins D, Emilio T, Souza PF, Levis C, Baccaro FB, Pinto JLPV, Moulatlet JM, Stark SC, Sarmento K, Araújo NO, Costa FRC, Schöngart J, Quesada CA, Saleska SR, Tomasella J, Magnusson WE. 2016. Forest structure along a 600 km transect of natural

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 24/26 disturbances and seasonality gradients in central-southern Amazonia. Journal of Ecology 104(5):1335–1346 DOI 10.1111/1365-2745.12596. Simões PI, Lima AP, Farias IP. 2010. The description of a cryptic species related to the pan Amazonian frog Allobates femoralis (Boulenger 1883) (Anura: Aromobatidae). Zootaxa 2406:1–28. Simões PI, Lima AP, Magnusson WE, Hödl W, Amézquita A. 2008. Acoustic and morphological differentiation in the frog Allobates femoralis: relationships with the upper Madeira River and other potential geological barriers. Biotropica 40(5):607–614 DOI 10.1111/j.1744-7429.2008.00416.x. Smith DHV, Jones B, Randall L, Prescott DRC. 2014. Difference in detection and occupancy between two anurans: the importance of species-speciﬁc monitoring. Herpetological Conservation and Biology 9:267–277. Smith MA, Green DM. 2005. Dispersal and the metapopulation paradigm in amphibian ecology and conservation: are all amphibian populations metapopulations? Ecography 28(1):110–128 DOI 10.1111/j.0906-7590.2005.04042.x. Sombroek W. 2000. Amazon landforms and soils in relation to biological diversity. Acta Amazonica 30(1):81 DOI 10.1590/1809-43922000301100. Sombroek W. 2001. Spatial and temporal patterns of Amazon rainfall. AMBIO: A Journal of the Human Environment 30(7):388–396 DOI 10.1579/0044-7447-30.7.388. Sousa TEL. 2007. Distribuição de palmeiras (Arecaceae) ao longo de gradientes ambientais no baixo interﬂúvio Purus-Madeira, Brasil. Dissertação de mestrado, INPA/UFAM. Stephens PA, Pettorelli N, Barlow J, Whittingham MJ, Cadotte MW. 2015. Management by proxy? The use of indices in applied ecology. Journal Applied Ecology 52(1):1–6 DOI 10.1111/1365-2664.12383. Stuart SN, Chanson JS, Cox NA, Young BE, Rodrigues ASL, Fischman DL, Waller RW. 2004. Status and trends of amphibian declines and extinctions worldwide. Science 306(5702):1783–1786 DOI 10.1126/science.1103538. Tews J, Brose U, Grimm V, Tielbörger K, Wichmann MC, Schwager M, Jeltsch F. 2004. Animal species diversity driven by habitat heterogeneity: the importance of keystone structures. Journal of Biogeography 31(1):79–92 DOI 10.1046/j.0305-0270.2003.00994.x. Tuomisto H, Ruokolainen K, Aguilas M, Sarmiento A. 2003. Floristic patterns along a 43 km long transect in Amazonian rain Forest. Journal of Ecology 91(5):743–756 DOI 10.1046/j.1365-2745.2003.00802.x. Tyre AJ, Tenhumberg B, Field SA, Niejalke D, Parris K, Possingham HP. 2003. Improving precision and reducing bias in biological surveys: estimating false-negative error rates. Ecological Applications 13(6):1790–1801 DOI 10.1890/02-5078. Ursprung E, Ringler M, Jehle R, Hödl W. 2011. Strong male/male competition allows for nonchoosy females: high levels of polygynandry in a territorial frog with paternal care. Molecular Ecology 20(8):1759–1771 DOI 10.1111/j.1365-294x.2011.05056.x. Warren DL. 2012. In defense of ‘niche modeling’. Trends in Ecology & Evolution 27(9):497–500 DOI 10.1016/j.tree.2012.03.010. Woinarski JCZ, Fisher A, Milne D. 1999. Distribution patterns of vertebrates in relation to an extensive rainfall gradient and variation in soil texture in the tropical savannas of the Northern Territory, Australia. Journal of Tropical Ecology 15(4):381–398 DOI 10.1017/s0266467499000905. Zimmerman BL. 1994. Audio strip transects. In: Heyer WR, Donnelly MA, McDiarmid RW, Hayek LAC, Foster MS, eds. Measuring and Monitoring Biological Diversity: Standard Methods for Amphibians. Washington, D.C.: Smithsonian Institution Press, 92–96.

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 25/26 Zimmerman BL, Bierregaard R. 1986. Relevance of the equilibrium theory of island biogeography and species-area relations to conservation with a case from Amazonia. Journal of Biogeography 13(2):133–143 DOI 10.2307/2844988. Zimmerman BL, Simberloff D. 1996. An historical interpretation of habitat use by frogs in a central Amazonian forest. Journal of Biogeography 23(1):27–46 DOI 10.1046/j.1365-2699.1996.d01-218.x. Zurell D, Thuiller W, Pagel J, Cabral JS, Münkemüller T, Gravel D, Dullinger S, Normand S, Schiffers KH, Moore KA, Zimmermann NE. 2016. Benchmarking novel approaches for modeling species range dynamics. Global Change Biology 22(8):2651–2664 DOI 10.1111/gcb.13251. Zuur A, Ieno EN, Walker N, Savellev AA, Smith GM. 2009. Mixed Effects Models and Extensions in Ecology with R. New York: Springer.

Ferreira et al. (2018), PeerJ, DOI 10.7717/peerj.5424 26/26